Explanation in artificial intelligence: Insights from the social sciences

Miller, Tim

doi:10.1016/j.artint.2018.07.007

Cited by 3,098 publications

(2,724 citation statements)

References 143 publications

Supporting

Mentioning

2,324

Contrasting

Unclassified

Order By: Relevance

“…First we provide some definitions to explain what kind of explainability we mean—this will lead us to the term “Causability” in contrast to the well‐known term “Causality”; then we discuss briefly the state‐of‐the‐art of some current explainable models, and continue with an example and a medical use‐case from histopathology. We conclude with pointing to the urgent need of a systems causability scale to measure the quality of an explanation (Hoffman, Mueller, Klein, & Litman, ), which must also include social aspects of human communication (Miller, ).…”

Section: Introductionmentioning

confidence: 99%

Causability and explainability of artificial intelligence in medicine

Holzinger

Langs

Denk

et al. 2019

WIREs Data Min & Knowl

907

571

View full text Add to dashboard Cite

Explainable artificial intelligence (AI) is attracting much interest in medicine. Technically, the problem of explainability is as old as AI itself and classic AI represented comprehensible retraceable approaches. However, their weakness was in dealing with uncertainties of the real world. Through the introduction of probabilistic learning, applications became increasingly successful, but increasingly opaque. Explainable AI deals with the implementation of transparency and traceability of statistical black‐box machine learning methods, particularly deep learning (DL). We argue that there is a need to go beyond explainable AI. To reach a level of explainable medicine we need causability. In the same way that usability encompasses measurements for the quality of use, causability encompasses measurements for the quality of explanations. In this article, we provide some necessary definitions to discriminate between explainability and causability as well as a use‐case of DL interpretation and of human explanation in histopathology. The main contribution of this article is the notion of causability, which is differentiated from explainability in that causability is a property of a person, while explainability is a property of a system This article is categorized under: Fundamental Concepts of Data and Knowledge > Human Centricity and User Interaction

show abstract

Section: Introductionmentioning

confidence: 99%

Causability and explainability of artificial intelligence in medicine

Holzinger

Langs

Denk

et al. 2019

WIREs Data Min & Knowl

907

571

View full text Add to dashboard Cite

show abstract

“…A major shortcoming of ML methods in general, and of DL methods in particular, is that the learned relations are hidden under very complicated prediction functions. However, recent years have seen the emergence of a whole field of ML called 'Explainable Artificial Intelligence' to face this issue (Miller 2019), and techniques and methodologies have been introduced to study what the ML/DL models are learning (Montavon et al 2018). In this work, we focus on an efficient method to explain the model predictions called regression activation mapping (RAM) (Zhou et al 2016, Wang and Yang 2017.…”

Section: Introductionmentioning

confidence: 99%

Estimating and understanding crop yields with explainable deep learning in the Indian Wheat Belt

Wolanin

Mateo-García

Camps‐Valls

et al. 2020

Environ. Res. Lett.

122

View full text Add to dashboard Cite

Forecasting crop yields is becoming increasingly important under the current context in which food security needs to be ensured despite the challenges brought by climate change, an expanding world population accompanied by rising incomes, increasing soil erosion, and decreasing water resources. Temperature, radiation, water availability and other environmental conditions influence crop growth, development, and final grain yield in a complex nonlinear manner. Machine learning (ML) techniques, and deep learning (DL) methods in particular, can account for such nonlinear relations between yield and its covariates. However, they typically lack transparency and interpretability, since the way the predictions are derived is not directly evident. Yet, in the context of yield forecasting, understanding which are the underlying factors behind both a predicted loss or gain is of great relevance. Here, we explore how to benefit from the increased predictive performance of DL methods while maintaining the ability to interpret how the models achieve their results. To do so, we applied a deep neural network to multivariate time series of vegetation and meteorological data to estimate the wheat yield in the Indian Wheat Belt. Then, we visualized and analyzed the features and yield drivers learned by the model with the use of regression activation maps. The DL model outperformed other tested models (ridge regression and random forest) and facilitated the interpretation of variables and processes that lead to yield variability. The learned features were mostly related to the length of the growing season, and temperature and light conditions during this time. For example, our results showed that high yields in 2012 were associated with low temperatures accompanied by sunny conditions during the growing period. The proposed methodology can be used for other crops and regions in order to facilitate application of DL models in agriculture.

show abstract

“…AI explanation is necessary to ensure trust (Lyons, ; Theodorou, Wortham, & Bryson, ) and is required by EU data protection laws. Explanation should be grounded in moral and social concepts, including values, social norms and relationships, commitments, habits, motives and goals (Miller, ).…”

Section: Ethical Pas: Design For Valuesmentioning

confidence: 99%

Supporting and challenging learners through pedagogical agents: Addressing ethical issues through designing for values

Richards¹,

Dignum²

2019

Brit J Educational Tech

View full text Add to dashboard Cite

Pedagogical Agents (PAs) that would guide interactions in intelligent learning environments were envisioned two decades ago. These early animated characters had been shown to deliver learning benefits. However, little was understood regarding what aspects were beneficial for learning and what sort of learning PAs were suitable for. This article considers the current and future use of PAs to support and challenge learners from three perspectives. Firstly, we look at PAs from a practical perspective to consider what Intelligent Virtual Agents are, the roles they play in education and beyond and the underlying technologies and theories driving them. Next we take a pedagogical perspective to consider the vision, pedagogical approaches supported and new possible uses of PAs. This leads us to the political perspective to consider the values, ethics and societal impacts of PAs. Drawing all three perspectives together we present a design for values approach to designing ethical and socially responsible PAs.

show abstract

Explanation in artificial intelligence: Insights from the social sciences

Cited by 3,098 publications

References 143 publications

Causability and explainability of artificial intelligence in medicine

Causability and explainability of artificial intelligence in medicine

Estimating and understanding crop yields with explainable deep learning in the Indian Wheat Belt

Supporting and challenging learners through pedagogical agents: Addressing ethical issues through designing for values

Contact Info

Product

Resources

About